Analysis from Unigestion
- Our core Q2 scenario is a -5% quarterly (non-annualized) decline in GDP growth across both the US and Europe.
- A liquidity squeeze is a key part of our tail-risk scenario, particularly in the wake of market indifference to monetary stimulus.
- Fiscal support is substantial across the US and Europe, but it is also uneven and presents a number of long-term risks.
- Valuations are attractive but we believe investors are underestimating the extent of the shock. We remain defensive.
The COVID-19 shock is one of the largest of the past 100 years. Although its impact is uncertain, it will be massive, and a rude awakening for investors lulled to sleep by accommodative central banks over the last decade. We think the extent of the shock is underestimated by many; prolonged shocks are difficult for investors to manage and can create feedback loops, and even with the combined global stimulus, we are likely to see choppy markets ahead.
The macro situation is very concerning, market sentiment is volatile, and valuation gaps in stocks and credit appear tempting. However, we think it is too soon to take on any meaningful risk at this point: we would need to see clear signs of an economic recovery first. Despite the large and coordinated stimulation from governments and central banks, growth is getting crushed by the virus’s impact. We believe that next quarter will be volatile with (1) growth assets likely exhibiting negative returns, (2) typical hedges showing disappointing performances, and (3) a tail risk that falling liquidity could accelerate the downturn in markets.
The core of our pessimism comes from our macro assessment: we think markets are broadly underestimating the impact on growth, earnings contraction and credit defaults, especially across the high yield space. The quarter should hopefully, see the end of the COVID-19 virus outbreak but not its impact on the economy and markets, and we remain on the defensive side.
Its not looking good
We have been communicating our growth scenario actively over the past month: we expect a sharp recession across the world. This growth contraction is hard to gauge because of the nature of this shock: it has been rapid, sudden and we lack the full spectrum of the usual statistics to measure its extent.
To overcome these handicaps and estimate its impact, we have investigated several leads, including
modeling the economic impact of the pandemic based on our in-house hypotheses, comparing the past three world recessions, and analyzing the situation in Asia. Across these different perspectives, we get a
consistent picture of a severe economic slowdown to which markets will react adversely.
The first analysis we conducted was a scenario analysis, investigating the potential impact of a prolonged consumption contraction. To understand the likely impact on the economy, we considered a model for a flu pandemic developed in the epidemiological literature (Health Outcomes and Costs of CommunityMitigation Strategies for an Influenza Pandemic in the United States, Perlroth et. al., Clin. Inf. Dis. 2010 Jan15;50(2):165-74). By looking at the decomposition of the US and Eurozone GDP and using a consumption-cut quarantine scenario, we can estimate the impact of the current social distancing measures on these two major economies. In quarterly terms, a quarantine that lasts for two months (as in China), with 90% of the population complying and cutting consumption by 70%, would result in a -5%quarter-on-quarter growth shock. This would bring 2020 US growth to -3% and Eurozone growth to -4%year on year, accounting for growth normalization in the second half of the year. Thus, according to our core scenario, 2020 should see the first global recession since the Great Financial Crisis (GFC). This
scenario assumes that the quarantine will have a limited impact on financial services, insurance, utilities
and healthcare services, and leaves 20% of usual energy consumption. If the current shock is not
temporary, investments will decline and lead to a more severe and longer-lasting scenario. This remains our tail-risk scenario, even though we are already seeing evidence of deteriorating investment
perspectives in Asia. In such a case, the -5% estimate could become as much as -8%, the largest macro
shock from such a short period ever seen in modern times . These scenarios are summarised in
Figure 1 below.
For comparison’s sake, during the fourth quarter of 2008 (the largest macro shock so far), US GDP
contracted by 1.5% quarter-on- quarter in non-annualised terms. We are currently expecting a one-off shock of about three times this magnitude. The COVID-19 crisis is therefore shaping up to be the sharpest decline in growth witnessed since World War II. However, it should be temporary, meaning that being able to track down its evolution on a daily basis is essential.
In addition, this crisis comes with a new challenge: the evolution of the underlying phenomenon is hard to measure. The contamination is highly dynamic on a daily basis, and so we have equipped ourselves with different indicators to understand its evolution: the Unigestion Newscaster and Nowcasters. TheUnigestion Growth Newscaster is an indicator that tracks public reporting on economic growth acrossthe world. This indicator is intended to complement our existing Growth Nowcaster by making use of the latest technologies in alternative data. Whereas our Growth Nowcaster assesses traditional macro data, the Growth Newscaster reads a vast array of news articles and summarises how the world is assessing the impact of any particular event on the broader economy. Hence it is expected to be faster-moving (at the cost of additional noise) than our Nowcaster but also more reactive, especially around turning points.
Since late January, after cresting at a neutral level, the Growth Newscaster has been on a one-way march downwards (see Figure 2). Initially, this was driven by news reports in the emerging world, specifically in China and the impact of the virus there. However, since mid-February, the deterioration has been driven by the developing world, as the virus evolved into a pandemic and news outlets turned their focus onto the economic fall-out. The indicator is now almost on par with the level at the onset of the GFC in late 2007 and in December 2018 amid fears of an impending recession. Figure 2 also compares the evolution of the Newscaster from its peak during these three periods. As the chart shows, the indicator has made most of the move down to GFC levels, suggesting the negative news stories today are on a par with those seen in late 2007. Given its nature, the Newscaster could see a quick snap-back if the news turns significantly less negative, but currently, both the level and trend are worrisome.
Recently, our nowcasting indicators have also started picking up a negative macro trend, with some lag.
Up to a couple of weeks ago, many cyclical surveys were still showing signs of improvement for the January-February period. Because this shock is exogenous, unlike all recessions except the oil shock of
1979, the entry into the recession was sudden. We therefore had to wait a couple weeks to collect the first data reflecting the extent of the situation. We now have this data, for Asia, Europe and the USA. Here are the results of our analyses:
The data in Asia shows a large and significant shock. In China, the shock seemed to be of a similar order of magnitude as the 2008 crisis up until March. The shock focuses more on services than on industry and affects demand, production expectations and, increasingly, the housing market. In SouthKorea, the shock is of a similar magnitude, with an additional feature: a collapse in investment perspective as future demand looks volatile and capacity production has decreased markedly, lowering the interest for CAPEX. Both these elements are illustrated in Figure 3. The daily estimates from our China Growth Nowcaster in March show the rapidity of the decline in economic conditions.The South Korea Growth Nowcaster highlights the significant and coincident decline in investment perspectives. From both cases, we learned two things: the deterioration can be fast and extraordinary.
When comparing the pace of the data deterioration in the US and in Europe, it is consistent with the worst moments of the past three recessions (1990, 2001 and 2008). Our World Growth Nowcaster declined by -0.3 standard deviations within a month, crossing the threshold separating expansion from recession on 27 March. Figure 4 shows historically the key moments when such a large change in a large aggregation of macro data last happened. Out of the eight cases, one is March 2020,another is October 1992. Apart from these cases, the rest of these rapid drops happened during the worst moments of the 1990 and 2008 recessions, two very significant world recessions. Such large and rapid declines in macro data are therefore rare and indicative of significant economic damage.
This move is very much in line with our expectation of a -5% quarterly (non-annualised) decline in GDP
growth for both the US and Europe. China’s fate is closely linked to these two, but the lack of transparency in its economy makes such a forecast more difficult. Nonetheless, it is hard to see how China could go unchallenged while its two largest clients experience such consumption contraction.
We have combined these different elements into a spectrum of scenarios that extrapolate the currently declining readings of our indicators:
- A ‘standard’ recession using the average evolution of our Nowcasters across the last three recessions
- A ‘2008-like’ scenario matching the current rapid deterioration in macro data to that in the fourth quarter of 2008,
- A ‘Newscaster’ scenario based on the deterioration seen in that indicator,
- An ‘Asian scenario’ that uses leading macro data from China, South Korea, and Taiwan,
- And finally a ‘pooled’ scenario averaging the above possibilities.
These scenarios are presented in Figure 5 below. In our view, this analytical approach is essential to deal with the unique features of the COVID-19 shock: gauging the extent of its impact requires reliable measures and we are firm believers in combining multiple reliable forecasts. Our current expectations are leaning toward the worst of these scenarios, inspired both by our Newscaster indicator and by the Asian experience. The recent intensity and speed of the decline in the US and European macro indicators also plays a significant role in this assessment.
Is enough help on the way?
Coordinated stimulation from fiscal and monetary authorities could help to alleviate part of the burden of this macro shock, so it is critical to assess these measures. Starting with fiscal policy, over the last month
we have seen governments in Europe and in the US implementing various fiscal measures in order to help
offset part of the impact of the lock-down. Two types of measures have been discussed: (1) direct spending of money to offset part of the losses in incomes and (2) funding loans to limit the number of defaulting businesses, especially small- and medium-sized ones that are most at risk.
Figure 6 compares the potential GDP destruction in our -5%/-8% pandemic simulation (illustrated in
Figure 1) expressed in billions of US dollars. It also compares those figures to what governments have been offering in terms of ‘direct support’ and ‘loans’. Our first conclusion is that in aggregate enough money has been gathered to face the needs arising from the core scenario of a 5% GDP contraction.
Indeed, between direct support and loans, more than enough has been pledged to deal with the roughly
USD 2.5 trillion damage that COVID-19 will likely bring to the world economy. However, the money
promised is extremely uneven from two angles: most of it is actually coming from the US (about USD 2tn)
while the rest of the proposed help is through loans, meaning that SMEs will bear the burden of the cost. We think this situation is not without risk: European SMEs would continue to struggle while the US
taxpayers will have to give away 1% of GDP growth for the next ten years to repay this 10% increase in
their debt to GDP.
Whichever way we look at it, we find limited comfort in this situation. While the aggregate amount of available money seems enough to overcome the needs of our core scenario, its imbalance creates a risk of failure that we think many may overlook. It also implies renewed central bank activism in order to prevent rates from increasing too much across the G10 world, as the fiscal premium required by investors to hold long-term bonds will likely increase with the rising debt of these countries.
Too early to be positive
In conclusion, when balancing the three sides of our dynamic investment process, we see a risk (the macro situation), a tail-risk (volatile and liquidity-dependant market sentiment) and a potentially deceptive set of opportunities (valuations). Our dynamic allocation entering into this quarter is defensive: our core scenario is that most people underestimate the consequences of the COVID-19 recession. This macro element is the number one reason for our pessimism: we think that markets are too optimistic in their perception of the impact of the lockdowns on the real economy. We notably expect earnings contraction to undershoot current, overly optimistic projections and believe high yield spreads to be at risk. This underestimation combined with sentiment could push markets down to lower levels, as the ability of central banks to prevent these drops is clearly in question. We expect to see better valuation opportunities ahead as we believe our macro scenario will be confirmed by the data. Over the upcoming quarter, we expect to keep an underweight to equities and high yield bonds, while being more constructive on investment-grade bonds for three reasons. First, we expect markets to show greater discrimination for higher quality assets at the expense of higher beta assets. Second, central bank asset purchase programs of investment grade bonds should support the market. Third, when combining both the previous elements, the risk-reward in investment grade markets has clearly improved. The failed attempts of the Fed to soothe markets and very low yields lead us to expect disappointing performance from government bonds should we see another round of market declines. We will become more constructive when we see three triggers: (1) an improved health/macro situation,(2) a convincing fiscal/monetary response, and (3) attractive valuations based on realistic expectations. None of these conditions are in place today, and so we remain defensive.